Delays and deficits in language are among the defining features of autism. Deficits in the domain of pragmatics, which involves the ability to use language in appropriate ways to communicate with others, are universally found in autism. Our ongoing research has shown that these pragmatic deficits are closely related to theory of mind impairments in children with autism. We have also found that among relatively high-functioning children with autism, about one-half demonstrate significant impairments in linguistic knowledge, and that these impairments mirror the profile of deficits that define specific language impairment (SLI). We hypothesize that this group of language impaired children with autism represents an important subtype that suggests overlap between these two disorders: autism and SLI. We plan to follow up these findings in experiments on language processing skills in children with autism, children with SLI and normal controls. Four series of experiments are proposed that investigate: (1) Phonological representations -deficits in non-word repetition and discrimination; (2) Morpho-syntax -deficits in grammatical knowledge related to marking tense using natural language samples, judgement and reaction-time tasks; (3) Prosody -the use of prosodic information in the speech signal to disambiguate syntactic information, or to convey a speaker?s emotional state; and (4) Word learning and naming -use of phonological or speaker?s intention cues to words presented in different contexts. We hypothesize that the subgroup of language impaired children with autism will perform like the SLI children in the experiments that tap phonological and grammatical knowledge. In contrast all the children with autism will perform worse that the SLI and normal controls on those experiments that entail inferring or using knowledge about speakers emotions or intentions. The findings from these studies will provide important new information about the language impairments in both autism and SLI and help to refine the language impaired subtype in autism for future genetic and neuroimaging studies.